Bone Morphogenetic Proteins 2 and 4 Are Selectively Expressed by Late Outgrowth Endothelial Progenitor Cells and Promote Neoangiogenesis
Objective— Endothelial progenitor cells are currently identified either by their surface antigen expression or by their generation of early colonies in culture (CFU-Hill). Another population, endothelial colony-forming cells (ECFCs), has strong vessel-forming capacity but is less well characterized. Given the potential usefulness of CFU-Hill and ECFCs as cell therapy products, their thorough characterization is of major importance.
Methods and Results— CFU-Hill and ECFCs were expanded from human cord and adult blood. Bone morphogenetic proteins 2 and 4 (BMP2/4) were selectively expressed by ECFCs but not by CFU-Hill. The BMP pathway was involved in ECFC commitment and angiogenic potential in vitro. In vivo, BMP inhibition strongly reduced plug vascularization in bFGF-containing Matrigel plugs implanted in C57/Bl6 mice. Moreover, ECFC exposure to BMP increased their therapeutic potential in a nude mouse model of hindlimb ischemia. In amputation specimens from patients with critical leg ischemia who had received a local therapeutic injection of bone marrow mononuclear cells, newly formed vessels were strongly positive for BMP2/4, suggesting that endothelial cells involved in neovascularization have an ECFC-like phenotype.
Conclusion— BMP2/4 are a marker of ECFCs and play a key role in ECFC commitment and outgrowth during neovascularization.
- endothelial progenitor cells
- bone morphogenetic protein
- critical leg ischemia
- cell therapy
Endothelial progenitor cells (EPCs) have been used for autologous angiogenic therapy and as a cardiovascular risk biomarker.1 Autologous mononuclear cells (MNCs) from bone marrow and peripheral blood have been tested for their angiogenic capacity in patients with ischemic diseases (for a review see1). However, the cell type responsible for the observed neovascularization and resulting clinical benefit is not known.
EPCs were initially defined as cells coexpressing the surface antigens CD34+, CD133+, and VEGFR2+. However, these antigens are also expressed on hematopoietic progenitor cells (HPCs), and two recent studies have shown that these cells give rise to hematopoietic cells in culture, but never to endothelial cells.2,3 Thus, EPC definition, quantification, and isolation is reaching a new level of complexity.
At least two populations of EPCs have been described.4 “Early” EPCs appear within 4 to 7 days of culture, are spindle-shaped, and express both endothelial and leukocyte markers. This cell population can be quantified as described by Hill.5 The number of these so-called “CFU-Hill” in peripheral blood has been proposed as a new biomarker, as it correlates negatively with cardiovascular risk factors.5 Endothelial colony-forming cells (ECFCs),6,7 also called “late” EPCs, develop after 2 to 3 weeks of culture and have the characteristics of precursor cells committed to the endothelial lineage. ECFCs are less well studied than “early” EPCs. By analogy with the terminology used for the hematopoietic lineage, the term colony forming unit-endothelial cells (CFU-ECs) should be restricted to endothelial colonies isolated in a 3D culture system such as methylcellulose.8,9 However, the term “endothelial progenitor cells” (EPCs) is controversial. Indeed, cells obtained in culture are already differentiated toward the endothelial lineage, and are thus referred to below as CFU-Hill and ECFCs rather than “progenitors”, based on the stringent criteria of Ingram and Yoder.10
To better characterize CFU-Hill and ECFCs, we examined their expression of selected embryonic stem cell genes. We found that bone morphogenetic proteins (BMPs) are selectively expressed by ECFCs but not by CFU-Hill. In addition, we show that the BMP pathway augments the expansion, proliferation, and migration of ECFCs, as well as their angiogenic potential in vivo.
Culture of ECFCs, CFU-Hill, and CFU-ECs
ECFC culture and characterization have been described in detail elsewhere1,11–15 and the cell culture method is described in supplemental materials, (available online at http://atvb.ahajournals.org). CFU-Hill were cultured using the EndoCult Liquid Medium kit (StemCell Technologies) according to the manufacturer’s instructions. Recombinant noggin from R&D systems was added to ECFCs and CFU-Hill cultures on the first day of culture and was present in all freshly added media. Gene expression profiles of ECFCs and CFU-Hill were compared with those of human embryonic stem cells (hESC), kindly provided by Dr Michel Pucéat (UMR861, Evry, France). For clonogenic assay of endothelial colonies (CFU-ECs), MNCs were seeded (4 or 8×105 cells per well) in basic methylcellulose Stem alpha 1 (Stem alpha) with 100 ng/mL human recombinant VEGF (R&D Systems).
In Vitro Angiogenic Assays
All experiments were performed in duplicate and repeated with cells from at least three different ECFCs, CFU-Hill, CFU-EC or CFU-GM colonies. These experiments are described in detail in supplemental materials.
In Vivo Matrigel Plug
These experiments are described in detail in supplemental materials.
In Vivo Quantification of ECFC-Induced Neovascularization
Nude mice underwent surgery to induce unilateral hindlimb ischemia as previously described16 and were then randomly assigned to injection of ECFCs treated with BMP2 or BMP4 for 48 hours, or untreated ECFCs. After 1 week, the ischemic/normal limb blood flow ratio was determined by using a laser Doppler perfusion imaging system (Moor Instruments). Vessel density was evaluated by using high-definition microangiography (Trophy system) after 14 days of treatment, as previously described.16
Local Injection of BM-MNCs in Patients With Critical Leg Ischemia
The Optimization of Progenitor Endothelial Cells in the Treatment of Critical Leg Ischemia (OPTIPEC) clinical trial is a multicenter, phase I nonrandomized study. Patients were eligible for the protocol if they had critical leg ischemia associated with limited gangrene or a nonhealing ischemic ulcer and if they were not eligible for surgical revascularization or percutaneous angioplasty, or if such a procedure had little chance of success. The cell therapy protocol was similar to that initially published by Tateishi-Yuyama et al.17 Four patients treated with BM-MNCs were studied here; their clinical characteristics are described elsewhere.18
BMP2/4 Are Selectively Expressed by ECFCs
CFU-Hill and ECFCs can be readily distinguished in culture on the basis of their morphological and clonogenic properties. To obtain CFU-Hill, MNCs were harvested and seeded onto fibronectin-coated tissue culture plates in Endocult medium. After 48 hours nonadherent cells were collected and replated in the same culture media for an additional 3 days on fibronectin-coated tissue culture plates. At day 5, CFU-Hill colonies were identified by the presence of elongated sprouting cells radiating from a central core of round cells (supplemental Figure IA). To obtain ECFCs, MNCs were harvested and seeded onto collagen-coated tissue culture plates in EGM2. In contrast to the CFU-Hill method, nonadherent cells were discarded daily and fresh EGM-2 medium was added to the cultures. ECFC colonies that originated from adherent cells (supplemental Figure IB) were identified between 7 and 30 days of culture. Differences in the culture methods and, particularly adherence/nonadherence to different substrates at early time points, suggest that ECFCs originate from an adherent cell and CFU-Hill from a nonadherent cell, and that the two cell types represent clonally distinct populations.10
However, selective markers are required to identify cells able to merge into neovessels. To better define ECFCs and CFU-Hill and to identify new markers differently expressed by these cells, we studied the expression of selected genes involved in stem cell biology by using real-time quantitative RT-PCR (Figure 1A). We compared their expression with that of human embryonic stem cells. We chose to explore the three classes of morphogens with distinct functions in early embryonic development, namely hedgehog proteins, wingless proteins, and bone morphogenetic proteins (BMPs). We also examined the expression of molecules with a role in hematopoietic stem cells (BMI-1), in the side population (ABCG-2), and finally Nanog and Oct4, which were recently shown to be expressed by a subset of EPCs.19 Numerous stem cell gene mRNA were found in both CFU-Hill and ECFCs (Figure 1A). The three representatives of the hedgehog family were quantified: SHH and IHH was poorly expressed by both cell types, and DHH was expressed at high levels only by ECFCs from cord blood. High levels of BMP2/4 mRNA and protein were detected in ECFCs derived from both cord blood and adult blood, whereas no expression was detected in CFU-Hill (Figure 1A and 1B). Moreover, BMP2/4 were also expressed at high levels in hESC (Figure 1A). As BMP2/4 expression distinguishes ECFCs from CFU-Hill, we restricted the rest of this study to BMP2/4.
CFU-EC Obtained in Methylcellulose Do Not Express BMP2 or -4
The methylcellulose assay can be used to isolate progenitor cells. Precursors and mature cells do not survive in the three-dimensional medium, whereas progenitors give rise to colony-forming units (CFU) that can be counted. When VEGF alone was added to methylcellulose,9 small round colonies, corresponding to CFU-ECs, were observed after 14 days of culture. They consisted of a cluster of round cells in the center and lengthened cells at the periphery (Figure 2A). The endothelial phenotype of CFU-ECs was shown by the expression of von Willebrand factor (vWF) (Figure 2B), and by their ability to incorporate Di-Ac-LDL (supplemental Figure IIA). These cells were harvested, and their endothelial and hematopoietic gene expression was studied at the mRNA level. Overall, CFU-ECs showed a CFU-Hill profile, with strong CD14 positivity (Figure 2C, and supplemental Figure IIB). We further showed that CFU-EC had features of proliferative cells, despite their CFU-Hill phenotype. Indeed, they strongly expressed Ki67, as observed with ECFCs and CFU-GM, whereas Ki67 expression was weak in CFU-Hill. This result was confirmed by the large proportion of cells in S-phase (supplemental Figure IIC). Importantly, CFU-ECs did not express BMP2 or 4. These results suggested that the cells obtained in methylcellulose alpha1/VEGF had the CFU-Hill phenotype but retained a proliferative capacity similar to that of CFU-GM. CFU-ECs can thus be considered as cells that derive from a population of circulating CD14+ endothelial progenitors.
BMP Inhibition by Noggin Inhibits the Emergence of ECFCs
Given the well-described role of BMP in regulating vasculogenesis, we investigated the effect of BMP pathway manipulation on the emergence of ECFC colonies. Noggin, a BMP antagonist that acts by binding to BMP, thereby inhibiting their binding to BMP receptors, significantly inhibited ECFC generation from cord and adult blood (respectively *P<0.0001 and *P=0.0003, Figure 3A). As expected, noggin had no effect on the number of CFU-Hill colonies, as counted with the Endocult system (P=0.78, supplemental Figure III).
BMP2/4 Activation Enhances Angiogenesis Mediated by ECFCs In Vitro
Flow cytometry showed that the three main BMP receptors (BMPR-1A, BMPR-1B, and BMPR-2) were expressed on more than 90% of ECFCs (supplemental Figure IV). Id genes are well-characterized BMP targets. We therefore studied the expression of the major Id gene (Id2) on BMP receptor activation in ECFCs. Id2 expression was induced by BMP2/4 treatment, and this effect was antagonized by noggin (supplemental Figure V).
To examine the effects of BMP2/4 on proliferation, ECFCs were treated with recombinant BMP2/4 for 72 hours. As shown in Figure 3B, BMP2 and BMP4 induced ECFC proliferation, the effect was significant because 1 ng/mL for BMP4 whereas it reached statistical significance only at 50 ng/mL for BMP2 (*P=0.008). Proliferation was also quantified by measuring H3 thymidine incorporation. BMP2 and 4 increased H3 thymidine incorporation, but only the effect of BMP4 reached significance (P=0.33 and 0.02 for BMP2 and BMP4, respectively; Figure 3C). A Boyden chamber model was used to test BMP chemoattraction of ECFCs. As shown in Figure 3D, a significant migratory effect of both BMP2/4 on ECFCs was observed (respectively *P=0.01 and *P=0.001). Finally, we examined the effects of 50 ng/mL BMPs on ECFC tube formation. In EBM2 medium (without any growth factors), few pseudotubes were observed in the absence of BMP. Both morphogens induced a strong and significant increase in tube formation, an effect that was inhibited by noggin (Figure 3E and 3F).
To examine the transcriptional effect of BMP activation, we used real-time quantitative RT-PCR to measure the expression of angiogenic factor genes previously described in EPCs,1 including VEGF-A, SDF-1, angiopoietin 2, and their respective receptors. No BMP-induced transcription of these genes was observed, ruling out an indirect proangiogenic effect involving these factors (supplemental Table).
Effect of BMP on Angiogenesis In Vivo
We firstly used the Matrigel plug assay as a model of in vivo angiogenesis. Basic-FGF-containing plugs, with and without noggin (850 ng per plug), were implanted subcutaneously in C57/B16 mice. The plugs were then recovered 14 days later and their blood vessel content was analyzed. Figure 4A clearly shows that the presence of noggin drastically reduced blood vessel formation. This effect was further quantified by measuring the hemoglobin content of the plugs, which was four times lower in the presence of noggin (*P<0.0001, supplemental Figure VI). To obtain direct evidence that BMP2/4 facilitate ECFC-mediated angiogenesis, we investigated the in vivo effects of BMP2/4 treatment on ECFCs. ECFCs stimulated with BMP2 or BMP4 were intravenously injected into nude mice that had undergone femoral artery ligature. As shown in Figure 4B, injection of ECFCs stimulated by BMP2 or BMP4 improved foot perfusion by 25% on day 7, as compared to injection of nonstimulated ECFCs (P=0.01 and 0.0002 respectively for BMP4 and BMP2). The mice were then killed on day 14 and a microangiographic score was calculated. Intravenous injection of BMP4-stimulated ECFCs significantly raised the ischemic/nonischemic limb angiography score by 40% compared to control animals injected with untreated ECFCs (P=0.04). The score was not improved by BMP2-treated ECFCs (P=0.01 compared to BMP4-treated ECFCs; Figure 4C). These results are in line with the stronger proliferation of BMP4- stimulated ECFCs compared to BMP2-stimulated ECFCs (Figure 3C).
Therapeutic Neovascularization by Bone Marrow Mononuclear Cells in Critical Leg Ischemia Patients Involves BMP2/4-Positive Endothelial Cells
Injection of bone marrow mononuclear cells into ischemic limbs increases collateral vessel formation. In the OPTIPEC trial, patients treated with BM-MNCs for critical leg ischemia had extensive neoangiogenesis in distal tissues.18 All the newly formed vessels were functional (blood cells were present within their lumen) and were limited by a continuous endothelium expressing the endothelial cell markers CD34, and vWF, proliferative marker Ki-67 (supplemental Figure VII), but not leukocyte antigen CD45 (Figure 5). Moreover, newly formed vessels formed after treatment with BM-MNCs are surrounded by vascular smooth muscle cells, indicating improved vessel maturation. Here, we found that these cells also expressed BMP2/4 (Figure 5), confirming that neovessels derive from the ECFC lineage in this setting. These data also strongly suggest that CFU-Hill, which does not express BMPs, do not participate directly in neovascularization.
Here we show that BMPs 2 and 4 are specifically expressed by ECFCs derived from adult and cord blood. BMPs are secreted growth factors that belong to the transforming growth factor (TGF) beta (TGF-β) superfamily and act via BMP type 1 (BMPR-1A and BMPR-1B), and type II2 (BMPR-2) serine/threonine kinase receptors. BMP2/4 appear to be the BMP group most strongly involved in vascular development.20
BMP2/4 Are Markers of ECFCs
We found that BMP2/4 are not expressed by CFU-Hill, contrary to ECFCs, whether these cells are derived from adult or cord blood, confirming the common ECFC lineage hypothesis forwarded by Ingram et al.6 We also show that the BMP pathway regulates ECFC commitment, as noggin addition to MNC cultures significantly attenuated ECFC growth. This suggests that human progenitor cell commitment to the endothelial lineage requires BMPs, as shown for the hemangioblast in ESC.21,22
Based on the method of quantification of hematopoietic progenitors CFU-GM by the clonogenic test in methylcellulose, we isolated progenitor cells from peripheral blood in the presence of VEGF and characterized the endothelial cell colony-forming units (CFU-ECs) thus obtained. This confirmed the monocytic origin of endothelial progenitors quantified by methylcellulose assay and suggested that BMPs are specific for the ECFC lineage. Indeed, we found that BMPs were expressed by ECFCs but not by cells with a monocytic phenotype cultured in 2D (CFU-Hill) or 3D systems (CFU-ECs). According to the markers tested (BMP, leukocyte, and endothelial markers), the genomic expression profile of CFU-ECs was similar to that of CFU-Hill. Other data must complete this characterization to evaluate the exact implication of CFU-EC obtained in 3D systems compared to CFU-Hill.
This work also suggests that BMPs, in addition to distinguishing CFU-Hill from ECFCs, are markers of immaturity. Indeed, BMP2/4 are expressed at high levels in human embryonic stem cells, as in ECFCs (Figure 1A). Among the markers tested on hESCs, only Nanog and Oct4 were expressed at a high level and were also expressed in a subpopulation of “early EPCs”, as described by Romagnani et al.19
BMP2/4 Enhances ECFC-Induced Angiogenesis
BMPs also act as proangiogenic factors, inducing ECFC proliferation, migration, and pseudotube formation in vitro and in vivo in a model of hindlimb ischemia. Recombinant BMP activated BMP receptors, as witnessed by Id2 gene activation, leading to increased proliferation, migration and pseudotube formation in vitro. Id has been shown to play an important role in mediating an angiogenic response during embryonic development and during post natal tumor angiogenesis.23
We cannot exclude the possibility that BMP2/4 promote angiogenesis by mechanisms other than direct ECFC stimulation. Indeed, BMP2/4 induce VEGF expression in osteoblasts.24 However, in our experimental conditions, BMP2/4 failed to stimulate the gene expression of VEGF, SDF-1 and angiopoietin 2 or their receptors in ECFCs (supplemental Table). Moreover, BMP2/4 may repress antiangiogenic factors. Indeed, another member of the BMP family, BMP1, has been implicated in the processing of a major basement membrane proteoglycan and in the release of an antiangiogenic factor.25 However, it is very difficult to measure at mRNA or protein level the expression of antiangiogenic factors because most of them are fragments of different molecules of the extracellular matrix: thrombospondin-1, endostatin, tumstatin, angiostatin, or 16-kDa N-terminal fragment of human prolactin. Other studies are needed to explore BMP interaction with other antiangiogenic factors that could be expressed or bound by ECFCs: CXCL4 (platelet factor 4), members of serpin superfamily like antithrombin or pigment epithelium derived factor (PEDF), or vasohibin. Indeed, such as observed with VEGF, BMP, could increase vasohibin synthesis, which functions as an endothelium-derived negative feedback regulator of angiogenesis.26 Finally, some BMP antagonists are probably secreted by ECFCs. Because BMPs have been shown to modulate synthesis of their own antagonist Noggin,27 we can hypothesize that modulation of BMP antagonist synthesis could have proangiogenic properties. BMP antagonists have been shown to activate endothelial cell receptors in a BMP-independent manner,28 thus BMP synthesis by ECFCs could also act by blocking potential antiangiogenic properties of these molecules.
BMP2/4 Are Expressed by Neovessels in Patients Who Receive Autologous BM-MNCs
We then examined whether BMP2/4 expression could be used to distinguish the origin of endothelial cells incorporated in human neovessels. We found that neovascularization observed in patients who had received autologous BM-MNCs by the intramuscular route involved cells expressing BMP2/4. However, it is unclear whether the highly proliferative cells contributing to neovessel formation in these conditions originate from the cell therapy product and/or from an endothelial “side population” of progenitors. These results are in line with those of Yoder et al, who found that ECFCs, contrary to CFU-Hill, formed functional human-murine chimeric vessels in a short-term xenograft model of blood vessel formation.7 Our clinical results suggest that neovascularization involves cells of the same lineage as ECFCs.
This study provides the first evidence that ECFCs express BMP2/4 morphogens. BMPs can be used as a marker of immaturity (also expressed by hESCs), as well as a marker of the ECFC lineage (not expressed by CFU-Hill), and finally as an angiogenic marker during ECFC commitment and expansion. Moreover, neovascularization observed in patients with critical leg ischemia treated with BM-MNCs involves BMP-expressing cells with an endothelial phenotype. This suggests that CFU-Hill, although able to promote angiogenesis through growth factor secretion, are not incorporated into functional neovessels in humans, as previously shown in mice. We propose a new classification for cells derived from EPCs in culture based on BMP2/4 expression: BMP-positive endothelial precursors correspond to ECFCs (late EPCs or BOECs), responsible for neovascularization, whereas BMP-negative endothelial precursors correspond to proangiogenic hematopoietic progenitor cells (CFU-Hill, early EPCs, circulating angiogenic cells, or CFU-EC).
We thank Dr Michel Pucéat from UMR861 (Evry, France) for kindly providing cDNA from hESC. We thank Céline Loinard, José Vilar, Christophe Avignon, Isabelle Martinez, and Chantal Mandet for their technical assistance.
Sources of Funding
We received research grants from INSERM, Programme Hospitalier de Recherche Clinique OPTIPEC (PHRC AOM 03 034, sponsor: AP-HP), and the Leducq TransAtlantic Network of Excellence on Atherothrombosis Research (Grant 04CVD01).
Original received April 18, 2008; final version accepted September 17, 2008.
Timmermans F, Van Hauwermeiren F, De Smedt M, Raedt R, Plasschaert F, De Buyzere ML, Gillebert TC, Plum J, Vandekerckhove B. Endothelial outgrowth cells are not derived from CD133+ cells or CD45+ hematopoietic precursors. Arterioscler Thromb Vasc Biol. 2007; 27: 1572–1579.
Hur J, Yoon CH, Kim HS, Choi JH, Kang HJ, Hwang KK, Oh BH, Lee MM, Park YB. Characterization of two types of endothelial progenitor cells and their different contributions to neovasculogenesis. Arterioscler Thromb Vasc Biol. 2004; 24: 288–293.
Ingram DA, Mead LE, Tanaka H, Meade V, Fenoglio A, Mortell K, Pollok K, Ferkowicz MJ, Gilley D, Yoder MC. Identification of a novel hierarchy of endothelial progenitor cells using human peripheral and umbilical cord blood. Blood. 2004; 104: 2752–2760.
Yoder MC, Mead LE, Prater D, Krier TR, Mroueh KN, Li F, Krasich R, Temm CJ, Prchal JT, Ingram DA. Re-defining endothelial progenitor cells via clonal analysis and hematopoietic stem/progenitor cell principals. Blood. 2006.
Gehling UM, Ergun S, Schumacher U, Wagener C, Pantel K, Otte M, Schuch G, Schafhausen P, Mende T, Kilic N, Kluge K, Schafer B, Hossfeld DK, Fiedler W. In vitro differentiation of endothelial cells from AC133-positive progenitor cells. Blood. 2000; 95: 3106–3112.
Urbich C, Heeschen C, Aicher A, Dernbach E, Zeiher AM, Dimmeler S. Relevance of monocytic features for neovascularization capacity of circulating endothelial progenitor cells. Circulation. 2003; 108: 2511–2516.
Yoder MC, Mead LE, Prater D, Krier TR, Mroueh KN, Li F, Krasich R, Temm CJ, Prchal JT, Ingram DA. Redefining endothelial progenitor cells via clonal analysis and hematopoietic stem/progenitor cell principals. Blood. 2007; 109: 1801–1809.
Smadja DM, Bieche I, Helley D, Laurendeau I, Simonin G, Muller L, Aiach M, Gaussem P. Increased VEGFR2 expression during human late endothelial progenitor cells expansion enhances in vitro angiogenesis with up-regulation of integrin alpha(6). J Cell Mol Med. 2007; 11: 1149–1161.
Smadja DM, Bieche I, Uzan G, Bompais H, Muller L, Boisson-Vidal C, Vidaud M, Aiach M, Gaussem P. PAR-1 activation on human late endothelial progenitor cells enhances angiogenesis in vitro with upregulation of the SDF-1/CXCR4 system. Arterioscler Thromb Vasc Biol. 2005; 25: 2321–2327.
Foubert P, Silvestre JS, Souttou B, Barateau V, Martin C, Ebrahimian TG, Lere-Dean C, Contreres JO, Sulpice E, Levy BI, Plouet J, Tobelem G, Le Ricousse-Roussanne S. PSGL-1-mediated activation of EphB4 increases the proangiogenic potential of endothelial progenitor cells. J Clin Invest. 2007; 117: 1527–1537.
Tateishi-Yuyama E, Matsubara H, Murohara T, Ikeda U, Shintani S, Masaki H, Amano K, Kishimoto Y, Yoshimoto K, Akashi H, Shimada K, Iwasaka T, Imaizumi T. Therapeutic angiogenesis for patients with limb ischaemia by autologous transplantation of bone-marrow cells: a pilot study and a randomised controlled trial. Lancet. 2002; 360: 427–435.
Romagnani P, Annunziato F, Liotta F, Lazzeri E, Mazzinghi B, Frosali F, Cosmi L, Maggi L, Lasagni L, Scheffold A, Kruger M, Dimmeler S, Marra F, Gensini G, Maggi E, Romagnani S. CD14+CD34low cells with stem cell phenotypic and functional features are the major source of circulating endothelial progenitors. Circ Res. 2005; 97: 314–322.
Kennedy M, D'Souza SL, Lynch-Kattman M, Schwantz S, Keller G. Development of the hemangioblast defines the onset of hematopoiesis in human ES cell differentiation cultures. Blood. 2007; 109: 2679–2687.
Gonzalez EM, Reed CC, Bix G, Fu J, Zhang Y, Gopalakrishnan B, Greenspan DS, Iozzo RV. BMP-1/Tolloid-like metalloproteases process endorepellin, the angiostatic C-terminal fragment of perlecan. J Biol Chem. 2005; 280: 7080–7087.
Stabile H, Mitola S, Moroni E, Belleri M, Nicoli S, Coltrini D, Peri F, Pessi A, Orsatti L, Talamo F, Castronovo V, Waltregny D, Cotelli F, Ribatti D, Presta M. Bone morphogenic protein antagonist Drm/gremlin is a novel proangiogenic factor. Blood. 2007; 109: 1834–1840.